JOM

https://doi.org/10.1007/s11837-021-05032-y
 2021 The Minerals, Metals & Materials Society

SOLID FREEFORM FABRICATION 2021

Microstructure and Deformation Behavior of Additively

Manufactured 17–4 Stainless Steel: Laser Powder Bed Fusion
vs. Laser Powder Directed Energy Deposition

P.D. NEZHADFAR,1,2 PAUL R. GRADL,3 SHUAI SHAO,1,2

and NIMA SHAMSAEI 1,2,4

1.—Department of Mechanical Engineering, Auburn University, Auburn, AL 36849, USA.

2.—National Center for Additive Manufacturing Excellence (NCAME), Auburn
University, Auburn, AL 36849, USA. 3.—Propulsion Department, NASA Marshall Space Flight
Center, Huntsville, AL 35812, USA. 4.—e-mail: shamsaei@auburn.edu

This study aims to compare the microstructure of 17–4 PH stainless steel (SS)
manufactured via laser powder bed fusion (L-PBF) and laser powder directed
energy deposition (LP-DED) in non-heat treated (NHT) and heat treated
conditions. In addition, the room-temperature tensile behavior of heat-treated
L-PBF and LP-DED 17–4 PH SS has been investigated and compared with
that of the wrought counterpart with the same heat treatment conditions. The
results show that the L-PBF specimens have a finer microstructure (ferrite +
lath martensite) than the LP-DED ones (massive ferrite + Widmanstätten
ferrite) in NHT condition. Electron backscatter diffraction analysis shows that
the L-PBF and LP-DED specimens have twin-based substructure lath
martensite after heat treatment. Despite the lower tensile strength of the LP-
DED specimens compared with the L-PBF ones, the ductility of peak-aged LP-
DED specimens was reduced due to the presence of the d-ferrite phase having
a significant plastic deformation incompatibility with the martensite.

Among the various AM processes, laser powder

INTRODUCTION
Additive manufacturing (AM) has transformed
the manufacturing process of structural parts in
various industries such as aerospace, automotive,
biomedical, defense, and nuclear. The layer-by-layer
and track-by-track nature of AM processes allows
the manufacture of near-net-shaped parts with
complex geometries, reduces the cost and lead
times, and facilitates the fabrication of highly
customized parts for specific applications (e.g., in
the medical field).1 There are various types of AM
techniques for metallic materials classified based on
feedstock form (e.g., powder versus wire), feeding
mechanism (e.g., powder bed versus blown powder),
and energy source (e.g., laser versus electron
beam).2
(Received September 30, 2021; accepted November 5, 2021)
bed fusion (L-PBF) and laser powder directed
energy deposition (LP-DED) are the most prominent
laser-based processes, which have been extensively
investigated in the literature.3 The L-PBF process
provides more freedom in designing near-net-
shaped parts with higher geometrical resolution
and precision than the LP-DED process; however,
L-PBF is typically limited to a single-powder feed-
stock.4 On the other hand, LP-DED offers better
compatibility for multi-powder feedstock, function-
ally graded components, and the fabrication of large
parts.1
One of the specific characteristics of the laser-
based AM processes is the high heating/cooling
rates during fabrication, resulting in a refined
microstructure compared to conventionally manu-
factured (CM) counterparts. Accordingly, the AM
materials often have static mechanical properties
comparable to, and in some cases outperforming,
their CM counterparts.5,6 The mechanical perfor-
mance of the AM materials is primarily influenced

by the material’s structure (i.e., grain structure,
texture, surface roughness, and defect structure),
which is impacted by the thermal history induced
during the manufacturing. The thermal history
itself is governed by the process parameters (i.e.,
powder, scan speed, etc.) and design parameters
(i.e., part size and geometry, time interval, etc.). In
addition to the above-mentioned influential param-
eters, the AM process technique (e.g., L-PBF, LP-
DED, etc.) itself may also cause variations in the
microstructural and mechanical properties of a
material system.
The LP-DED process, due to the slower moving
melt pool (typically 10–30 mm/s7 compared to
1000–1700 mm/s for L-PBF8) created by a high
powered laser (typically 1000 W), has a consider-
ably lower cooling rate (by approximately three
orders of magnitude) than L-PBF,9,10 giving the
fabricated materials a distinct microstructure (i.e.,
grain structure, crystallographic texture, precipita-
tions, etc.) and defect content. Amato et al.11
reported the presence of a spheroidal/ellipsoidal c‘‘
in c matrix of non-heat treated (NHT) L-PBF IN718,
while Laves phase has been characterized in an
interdendritic c matrix of NHT LP-DED counter-
parts12,13 Xu et al.14 reported a lamellar a+ß struc-
ture in L-PBF Ti-6Al-4V, whereas Caroll et al.15
showed a Widmanstätten structure along with
grain boundary a phase in the coarse columnar
prior b grains in LP-DED Ti-6Al-4V. It has been
reported that even post-thermal heat treatment
may not completely diminish the differences in the
microstructure of the L-PBF and LP-DED AM
materials. Schneider et al.16 reported different
grain structures (i.e., size and morphology) for L-
PBF and LP-DED IN718 even after applying a
similar heat treatment procedure, which was attrib-
uted to the variation in their initial microstructure
(i.e., NHT condition). The L-PBF specimens showed
the highest degree of homogenization and more
refined grains than the LP-DED ones.
The microstructure variations of the L-PBF and
LP-DED AM materials result in different mechan-
ical responses. Babuska et al.4 recently reported a
higher tensile strength for L-PBF CoCr specimens
than for LP-DED counterparts. They attributed this
to the finer microstructure of the L-PBF specimens
due to the higher cooling rate compared with the
LP-DED ones. Donate-Buendia et al.17 compared
the microstructure, and mechanical properties of
oxide dispersion strengthened (ODS) steels manu-
factured via L-PBF and LP-DED processes. Com-
pared to LP-DED, the L-PBF process produced a
Table I. Chemical composition of 17-4 PH SS powders used for fabrication
C Cr Ni Cu Mn Si
L-PBF (Wt.%) 0.01 15.80 4.60 3.67 0.51 0.32
LP-DED (Wt.%) 0.01 16.39 4.17 3.32 0.06 0.78
Nezhadfar, Gradl, Shao, and Shamsaei
finer grain structure, reduced the agglomeration of
the strengthening particles (i.e., Y2O3), and yielded
higher hardness values in the ODS steel.
The significant dependence of the microstructure
and mechanical properties of such known materials
as IN718, Ti-6Al-4V, ODS steels, etc. on AM pro-
cesses necessitates the careful evaluation of this
dependence before mass adoption of AM to any
other materials in key industrial sectors. The 17-4
precipitation hardening (PH) stainless steel (SS) is
one of the most commonly used materials in various
industries (e.g., aerospace, energy, food processing,
etc.) due to its favorable chemical and mechanical
properties (e.g., high corrosion resistance, high
strength, and ductility, etc.).18 It has recently
drawn much attention from the AM community
due to its superior weldability.19,20 The sensitivity of
its microstructure to the cooling rate during solid-
ification, as well as the subsequent thermal expo-
sures, can potentially make the properties of 17–4
PH SS process-dependent. This study investigates
and compares the microstructure and room-temper-
ature tensile behavior of 17–4 PH SS manufactured
by the L-PBF and LP-DED processes. The grain
structure and phase constituent of the L-PBF and
LP-DED 17–4 PH SS specimens are characterized
and compared in the NHT condition. In addition,
the effect of various heat treatment conditions on
the crystallographic texture and room-temperature
tensile behavior of the L-PBF and LP-DED 17–4 PH
SS specimens is investigated.
EXPERIMENTAL
Two batches of argon-atomized 17–4 PH SS
powder were used to fabricate the specimens via
the L-PBF and LP-DED processes, respectively. The
detailed chemical compositions of the 17–4 PH SS
powders used for each AM process are listed in
Table I.
An EOS M290 machine was used to fabricate the
L-PBF cylindrical bars (/11 mm). The EOS default
process parameters, laser power of 220 W, scan
speed of 755 mm/s, hatching distance of 0.1 mm, and
layer thickness of 0.04 mm, were employed, and
nitrogen was used as the shielding gas. The LP-
DED cylindrical bars (/15.24 mm) were fabricated
using an RPM Innovations (RPMI) 557 machine.
The process parameters were: power of 1070 W,
layer height of 0.38 mm, travel speed of 1,016 mm/
min, and powder feed rate of 15.1 g/min, with argon
used as the shielding gas.
Nb Mo N O P S Nb+Ta Fe
0.32 0.21 0.02 0.04 0.035 0.014 0.27 Bal.
0.27 0.04 0.01 0.02 0.00 0.00 0.27 Bal.
Microstructure and Deformation Behavior of Additively Manufactured 17–4 Stainless Steel:
Laser Powder Bed Fusion vs. Laser Powder Directed Energy Deposition
The Thermo-Calc. software was used to generate
the Fe-Cr binary phase diagram to analyze the
phase transformations possible for 17-4 PH SS. The
TCFE9 thermodynamic database for various Fe-
based alloys and steels such as stainless steels was
employed.21,22 In addition, to incorporate the influ-
ence of all the alloy elements in generating the
phase diagram, the Nieq (nickel equivalent) and Creq
(chromium equivalent) were obtained following the
Schaeffler equations:23
Nieq ðwt%Þ ¼ %Ni þ 0:5ð%MnÞ þ 0:3ð%CuÞ
þ 25ð%NÞ þ 30ð%CÞ
Creq ðwt%Þ ¼ %Cr þ 2ð%SiÞ þ 1:5ð%MoÞ
þ 1:75ð%NbÞ
The Nieq and Creq values were further calculated
for the L-PBF to be 7.6 wt% and 16.6 wt%, respec-
tively. For the LP-DED 17-4 PH SS, the Nieq and
Creq values were 6.1 and 17.7 wt%, respectively.
Figure 1 shows the binary phase diagram generated
by Thermo-Calc. software; the Cr amounts for the L-
PBF and LP-DED specimens are indicated on the
phase diagram to predict the phase constituents at
various temperatures. According to the phase dia-
gram, the stress relieve (SR) heat treatment was
carried out at 650C/1 h for the L-PBF and LP-DED
specimens. It has been reported in Ref. 24 that this
SR procedure changes neither the microstructure
nor the mechanical properties of the L-PBF and LP-
Fig. 1. Ni-Cr binary phase diagram samples generated by Thermo-Calc. software using the TCFE9 thermodynamic database22.
DED 17-4 PH SS. All the L-PBF and LP-DED
specimens underwent hot isostatic pressing (HIP) at
1163C/3 h under 103 MPa pressure and were
solution (Sol) treated at 1050C/0.5 h [i.e., Condition
A (CA)] followed by air cooling. Some specimens for
each type (i.e., L-PBF and LP-DED specimens) were
aged at 482C/1 h (i.e., H900) known as peak-age,
and some were over-aged at 621C/4 h (i.e., H1150).
In the following text, the heat-treated specimens
(i.e., SR + HIP + Sol + Age) are dubbed CA-H900
and CA-H1150. After the full heat treatment, all the
specimens were machined to the final tensile testing
ð1Þ geometry following the ASTM E8 standard.25
The L-PBF and LP-DED specimens were cut in
the gage section transversely parallel to the build
direction for microstructure characterization prior
ð2Þ to tensile testing. The microstructure was charac-
terized in NHT and fully heat treated conditions.
The samples were ground and polished using sand-
papers with grits 320–4000 followed by a mirror-
finish polishing step using Chemo-met along with a
0.2-lm colloidal silica suspension. The microstruc-
ture was further characterized on the plane parallel
(XZ-plane) to the build direction using a Zeiss
Crossbeam 550 scanning electron microscope
(SEM) with an electron backscatter diffraction
(EBSD) detector. The specimens were mirror-fin-
ished using a vibratory polisher for 2 h before
conducting the EBSD analysis and electron chan-
neling contrast imaging (ECCI).

The tensile tests were conducted on the heat-
treated L-PBF and LP-DED 17-4 PH SS specimens
at room temperature at 0.005 mm/mm/min strain
rate. The fracture surfaces were further cleaned in a
bath of isopropanol and water using an ultrasonic
cleaner before performing fractography using SEM.
RESULTS AND DISCUSSION
NHT Microstructure
The NHT microstructures of the L-PBF and LP-
DED 17-4 PH SS specimens characterized via EBSD
analysis and ECCI are shown in Fig. 2. It is evident
that the L-PBF 17-4 PH SS specimen (Fig. 2a) has a
more refined grain structure as compared to the LP-
DED 17–4 PH SS one (Fig. 2b); the L-PBF specimen
consists of fine, primarily equiaxed, ferrite grains,
whereas the LP-DED one consists of very coarse
columnar ferrite grains. The ECCI images with
higher magnifications reveal that the microstruc-
ture of the L-PBF specimen consists of ferrite and
lath martensite. In contrast, the LP-DED specimen
has a ferritic microstructure consisting of massive
ferrite grains with Widmanstätten ferrites decorat-
ing the grain boundaries.
Fig. 2. Inverse pole figure (IPF) maps along the Z direction and ECCI micrographs of (a) L-PBF, and (b) LP-DED 17-4 PH SS specimens in NHT
condition.
Nezhadfar, Gradl, Shao, and Shamsaei
Since the martensite start temperature (i.e., Ms)
is above room temperature for the 17–4 PH SS
(Ms = 100–150C),26,27 the austenite formed during
solidification transforms to the martensite upon
cooling to room temperature. However, the ferritic
microstructure of the NHT AM 17-4 PH SS speci-
mens, regardless of the manufacturing technique,
has been attributed to the ‘‘austenite by-passing’’
mechanism. In this mechanism, the ferrite phase
formed from the liquid will not transfer to the
austenite due to the high cooling rate in AM
processes as compared to conventional manufactur-
ing techniques (e.g., casting, forging, etc.).28
The lath martensite in the microstructure of L-
PBF 17-4 PH SS specimen (see Fig. 2a), however,
indicates that the austenite has been formed during
the solidification. This may be attributed to the N2
shielding gas used to fabricate the L-PBF speci-
mens. It has been reported that N2 as the shielding
gas refines the grain structure and also stimulates
austenite formation during solidification.20,29,30
Therefore, the refined ferrite grains (average grain
size of 6 lm) and the lath martensite in the
microstructure are the result of using N2 as shield-
ing gas. The coarse microstructure of the LP-DED
Microstructure and Deformation Behavior of Additively Manufactured 17–4 Stainless Steel:
Laser Powder Bed Fusion vs. Laser Powder Directed Energy Deposition

17-4 PH SS specimen (average grain size of 93.1 Heat Treated Microstructure

lm) may be attributed to the lower cooling rate
The microstructure of the L-PBF and LP-DED 17-
induced in the LP-DED process as compared to the
4 PH SS specimens are compared in Fig. 3 for the
L-PBF one. Although the cooling rate of the LP-
CA-H900 and CA-H1150 heat treatment conditions.
DED process is believed to be above the threshold
Although there is a visible difference between the
for austenite by-passing, it has been reported to
microstructures of the NHT L-PBF and LP-DED 17-
have almost a two to three orders of magnitude
4 PH SS specimens (see Fig. 2), post-thermal treat-
slower cooling rate than the L-PBF process.4 In
ment (i.e., SR + HIP + HT) have evidently altered
addition, the ferrite phase is more stabilized in LP-
the microstructures to martensite-dominated ones
DED specimens due to the higher Creq/Nieq value for
for both the L-PBF and LP-DED specimens (see
the LP-DED 17-4 PH SS (2.9) compared to that of
Fig. 3). It can be seen in Fig. 3a and c that both the
the L-PBF counterpart (2.1). The significantly
L-PBF and LP-DED have a primarily martensitic
different NHT microstructures of the L-PBF and
microstructure in the CA-H900 heat treatment
LP-DED 17-4 PH SS specimens may not only lead to
condition with a minimal fraction of retained
drastically different mechanical properties but also
austenite (0.1%). The fraction of retained austen-
may be inferior in strength due to the absence of
ite is increased for both the L-PBF (Fig. 3b) and LP-
precipitates. In hope to resolve these issues, post-
DED (Fig. 3d) specimens that have undergone CA-
thermal treatments are typically performed to
H1150 heat treatment. This is due to the diffusion of
enhance the mechanical properties of the material,
austenite stabilizer elements (i.e., Cu, Ni, N, etc.) to
as well as reducing the differences in the
the grain boundaries during the long-term aging
microstructures of the L-PBF and LP-DED
(i.e., 4 h) at a temperature close to the austenite
specimens.
reversion (see Fig. 1), which may have resulted in
the nucleation of austenite grains.31 However,
depending on the size of the reverted austenite
grains, they may or may not transform to

Fig. 3. IPF maps (along the Z direction) and phase maps for the heat treated 17-4 PH SS specimens: (a) L-PBF and (c) LP-DED at CA-H900,
and (b) L-PBF and (d) LP-DED at CA-H1150. Note that the black boundaries in the magnified IPF maps (in the middle) represent the prior
austenite grain boundaries.

martensite upon cooling. It has been well estab-
lished that martensitic transformations are more
difficult for finer austenite grains, which may be
retained after heat treatment.32
As seen in Fig. 3b for the CA-H1150-treated L-
PBF specimen, the fraction of austenite (4.3%)
retained in the microstructure is higher than that of
the CA-H1150 LP-DED counterpart (0.5%), shown
in Fig. 3d. This can be partially ascribed to the finer
initial microstructure of the L-PBF specimen com-
pared with that of the LP-DED one, which results in
finer reverted austenite grains that are more diffi-
cult to transform to martensite upon cooling. More-
over, using N2 shielding gas for fabrication may
increase the austenite stabilization and hinder the
martensitic transformation. It can be observed that
the d-ferrite phase is retained in the microstructure
of the LP-DED specimens (see Fig. 3c and d),
regardless of the heat treatment condition, whereas
the d-ferrite phase is absent in the microstructure of
the N2-shielded L-PBF specimens. The N2 shielding
gas has been reported to reduce the d-ferrite phase
fraction retained in the microstructure of laser-
welded materials at room temperature. N as an
austenite stabilizer results in the transformation of
the d-ferrite phase to austenite during the solidifi-
cation, which reduces the fraction of d-ferrite phase
retained in the microstructure at room temperature.
Moreover, the L-PBF specimen has a lower Creq/
Nieq value than the LP-DED one, reducing ferrite
stability during solidification.
Texture Analysis
The variation in thermal history induced via L-
PBF and LP-DED processes influences the crystal-
lographic texture of the AM 17–4 PH SS specimens.
It has been shown that the NHT L-PBF 17–4 PH SS
specimen possesses a strong cube along with weak c-
fiber texture components, whereas no specific tex-
ture component has been reported for the NHT LP-
DED 17-4 PH SS specimen due to its very large
grain structure (see Fig. 2b).24 In addition, it has
been reported that performing SR heat treatment
assists with the recrystallization upon further heat-
treatment procedures and weakens the texture.24
To understand and compare the crystallographic
textures of the L-PBF and LP-DED 17–4 PH SS
specimens after applying full heat treatments (i.e.,
SR + HIP + Sol + Age), their corresponding orien-
tation distribution function (ODF) maps for each
condition are presented in Figs. 4 and 5. The ODF
maps have been analyzed in parallel to the build
and loading direction (i.e., Z) using a spherical
harmonics method, keeping the Euler angle 3 (i.e.,
u2) constant. The quantified fraction of the essential
texture components, i.e., Cube (C), Rotated Cube
(RC), Copper (Cu), Transformed Copper (TC), Goss
(G), Transformed Goss (TG), Brass (B), Trans-
formed Brass (TB), and c-fiber) are presented and
Nezhadfar, Gradl, Shao, and Shamsaei
compared for the L-PBF and LP-DED specimens in
both CA-H900 and CA-H1150 conditions.
In addition to the ODF maps (0 £ U1, U, u2 £ 90),
the fraction of texture components for the L-PBF
and LP-DED specimens which have undergone CA-
H900 and CA-H1150 heat treatments are also
provided in Figs. 4c and 5c, respectively. In the
CA-H900 heat treatment condition, the L-PBF
specimen (see Fig. 4a) has a lower texture intensity
[i.e., multiple random density (mrd)], mrdmax =
1.85, than the LP-DED one, mrdmax = 3.25 (see
Fig. 4b). This is likely due to the finer grain
structure of the L-PBF specimen and most likely
with more randomly oriented grains than LP-DED
one. The texture intensities for both the L-PBF
and LP-DED specimens are reduced as compared
to their NHT conditions ((mrdmax)L-PBF = 2.03,
(mrdmax)LP-DED = 20.87) reported in24 and become
more randomized. This is ascribed to the recrystal-
lization occurring during HIP and to further heat
treatment procedures (i.e., Sol + Age). Therefore,
full heat treatment could to some extent weaken
texture differences between the L-PBF and LP-DED
17–4 PH SS specimens induced due to the differ-
ences in the cooling rates of the AM processes.
The main texture component for the L-PBF
specimen at the CA-H900 heat treatment condition
is TB, whereas TC1 is the main texture component
for the LP-DED specimen (see Fig. 4c). It can be
seen in Fig. 5c that the conducting CA-H1150 heat
treatment will result in a more similar trend of
texture component fractions in the L-PBF and LP-
DED specimens as compared to the CA-H900 heat-
treated ones. This may be due to the longer duration
of aging close to the austenite reversion tempera-
ture, and likely have resulted in partial reversion of
the austenite. The reverted austenite then trans-
forms to the martensite. The transformed-type
texture components represent the texture formed
from the parent austenite phase (face-centered
cubic) during martensitic transformation.
Ping et al.33 have shown the formation mecha-
nism of lath martensite with a twin substructure
with a {112}<111> relationship. Figure 6 shows the
{112} pole figures as well as the linear misorienta-
tion profiles along the arrows elongated through the
selected prior austenite grain (PAG) for the L-PBF
and LP-DED specimens which have undergone CA-
H900 and CA-H1150. The poles coincidence in the
{112} maps for all the conditions are in line with the
reported pole figures for martensite with the Kur-
djumov–Sachs (K-S) relationship with the parent
austenite.34 In addition, the point-to-point misori-
entation profiles show the boundaries with 60
misorientation indicating the twin boundaries,35
which further proves the twin-based martensitic
transformation in the L-PBF and LP-DED 17-4 PH
SS specimens.
Microstructure and Deformation Behavior of Additively Manufactured 17–4 Stainless Steel:
Laser Powder Bed Fusion vs. Laser Powder Directed Energy Deposition

Fig. 4. ODF maps for the (a) L-PBF and (b) LP-DED 17-4 PH SS specimens, and (c) their quantified texture components for the CA-H900 heat
treatment condition. A schematic of the important texture components in body-centered cubic for u2 = 45 is also shown.

Tensile Behavior and Fractography Analysis (i.e., CA-H1150) significantly reduces the strength
and increases the ductility. The high tensile
The engineering stress-engineering strain curves
strength (i.e., Sy, Su) after the CA-H900 condition
for the L-PBF and LP-DED specimens which have
is attributed to the formation of nano-sized Cu-
undergone CA-H900 and CA-H1150 heat treatment
enriched precipitates in this heat treatment condi-
conditions are presented in Fig. 7a. The tensile
tion, while these precipitates are coarsened upon
properties (i.e., Sy, Su, and %El) for the L-PBF, LP-
over-aging resulting in higher ductility.36
DED, and wrought 17-4 PH SS are compared in
Comparing the L-PBF and LP-DED specimens,
Fig. 7b. In general, the peak-age heat treatment
the L-PBF specimens have higher tensile strength
(i.e., CA-H900) results in the highest tensile
than those of the LP-DED counterparts regardless
strength in 17–4 PH SS at the expense of the
of the heat treatment condition. This is because of
ductility, whereas the over-aging heat treatment
the finer microstructure obtained for the L-PBF
 Nezhadfar, Gradl, Shao, and Shamsaei

Fig. 5. ODF maps for the (a) L-PBF and (b) LP-DED 17-4 PH SS specimens, and (c) the quantified texture components for the CA-H1150 heat
treatment condition. A schematic illustration of the important texture components in body-centered cubic for u2 = 45 is also shown.

specimens compared with that of the LP-DED ones
(see Fig. 3). The L-PBF specimen has higher tensile
strength, whereas the LP-DED specimen possesses
comparable tensile strength and ductility compared
with the wrought 17-4 PH SS. The higher tensile
properties of the L-PBF 17-4 PH SS are ascribed to
its finer grain structure.
Although the CA-H900 heat treated LP-DED
specimen has a lower tensile strength than the L-
PBF counterpart and is expected to have higher
ductility, its ductility is comparable to that of the L-
PBF specimen with higher tensile strength. The loss
of the expected ductility in the CA-H900 heat-
treated LP-DED specimen is due to the presence of
d-ferrite in its microstructure (see Fig. 3c). Figure 8
shows the kernel average misorientation (KAM),
and the Schmid factor map of the body-centered
cubic slip system (i.e., {110} <111>) along the
loading direction (i.e., parallel to the Z axis) for
the LP-DED 17-4 PH SS specimens heat treated at
CA-H900 (Fig. 8a and b). The d-ferrite phase is
indicated by arrows in the KAM images and circled
in the Schmid factor maps. As shown in Fig. 8a and
b for the CA-H900 heat-treatment condition, the d-
Microstructure and Deformation Behavior of Additively Manufactured 17–4 Stainless Steel:
Laser Powder Bed Fusion vs. Laser Powder Directed Energy Deposition

Fig. 6. IPF maps, {112} pole figures, and the point-to-point misorientation maps showing the twin-based substructure of the martensite in
selected PAG of the L-PBF and LP-DED 17-4 PH SS specimens. CA-H900 condition: (a) L-PBF, (b) LP-DED, and CA-H1150 condition: (c) L-
PBF, and (d) LP-DED.

ferrite phase not only has well-defined edges but it
also has a strong contrast in the Schmid factor with
the martensitic matrix; in other words, there is a
high incompatibility between the d-ferrite and the
matrix resulting in loss of ductility considering that
the material possesses high strength.
The fracture surfaces of the L-PBF and LP-DED
17-4 PH SS specimens are compared in Figure 9a
for CA-H900 and Fig. 9b for the CA-H1150 heat-
treatment conditions. Both conditions result in
mostly axis-symmetric fracture surfaces with the
shear lips of CA-H900 condition occupying very
small area fractions and those of CA-H1150
 Nezhadfar, Gradl, Shao, and Shamsaei

Fig. 7. Tensile behaviors of the L-PBF and LP-DED 17-4 PH SS specimens which have undergone CA-H900 and CA-H1150 heat treatment
conditions: (a) flow stress curves, (b) summarized tensile properties, and (c) %elongation versus ultimate tensile strength. The wrought data has
been. taken from ASTM A69337

Fig. 8. Strain analysis of the d-ferrite phase for the LP-DED 17-4 PH SS specimen which has undergone CA-H900 heat treatment condition: (a)
KAM, (b) Schmid factor maps. Note the loading direction parallel to the Z-axis considered for generating the Schmid factor map.

condition being large. The very well-defined cup-
and-cone feature CA-H1150 fracture surfaces
(Fig. 9b) suggests that their fracture was very
ductile. This is consistent with the presence of
quasi-cleavage facets and cracks on the fracture
surfaces of the CA-H900 heat treated specimens
showing brittle behavior, while enormous dimples
formed on the fracture surfaces of the CA-H1150
heat treated specimens representing ductile
behavior.
Another main difference between the fracture
surfaces of the L-PBF and LP-DED specimens is the
presence of cracks in the CA-H900 condition, and
dimples in the CA-H1150 condition. This suggests
that the fracture surfaces of the former condition
had radial zones which were evidence of rapid crack
propagation, while the fractures of the latter condi-
tion were governed mainly by the nucleation,
growth, and coalescence of voids as well as final
shearing.38 The CA-H900-treated LP-DED speci-
men has larger cracks and facets than the L-PBF
ones, which resulted in the loss of ductility despite
of having lower strength. This may be related to the
presence of large d-ferrite in the microstructure of
the LP-DED specimen, having high incompatibility
with the matrix, which resulted in crack initiation
and growth along the lath martensite boundaries.
On the other hand, CA-H1150 treated LP-DED
specimens have larger and deeper dimples due to
the coarser microstructure, justifying its higher
ductility compared with the L-PBF ones.
CONCLUSION
This study characterized and compared the non-
heat treated (NHT) microstructures of 17–4 PH SS
specimens fabricated via L-PBF and LP-DED AM
processes. In addition, the effect of different heat-
Microstructure and Deformation Behavior of Additively Manufactured 17–4 Stainless Steel:
Laser Powder Bed Fusion vs. Laser Powder Directed Energy Deposition

Fig. 9. Tensile fracture surfaces of the L-PBF and LP-DED 17-4 PH SS specimens: (a) CA-H900, and (b) CA-H1150.

treatment procedures on the crystallographic tex-
ture and room-temperature tensile deformation
behavior of L-PBF and LP-DED specimens were
investigated. The following conclusions are drawn:
1. The NHT L-PBF 17–4 PH SS specimen had a
more refined microstructure, constituted of
equiaxed ferrite grains and lath martensite,
whereas the microstructure of the LP-DED
counterpart was composed of coarse massive
ferrite, with Widmanstätten ferrite grains dec-
orating the grain boundaries. The variation in
the cooling/solidification rates induced by the L-
PBF and LP-DED processes and the differences
in Creq/Nieq values for the L-PBF and LP-DED
specimens may have caused these differences in
their NHT microstructures.
2. The d-ferrite phase was retained in the
microstructure of the LP-DED 17–4 PH SS

specimen at room temperature. This was as-
cribed to the lower cooling/solidification rate in
the LP-DED process and higher Creq/Nieq value
of the LP-DED specimen compared to the L-PBF
one, which stabilized ferrite in the microstruc-
ture.
3. The over-aging heat treatment procedure (i.e.,
CA-H1150) likely resulted in the reversion of
austenite from martensite. However, the re-
verted austenite was retained in the microstruc-
ture of the L-PBF specimen, while it was
transformed to martensite in the LP-DED coun-
terpart upon cooling. The finer microstructure
in the L-PBF specimen resulted in finer reverted
austenite, which hindered the martensitic
transformation.
4. A full heat treatment cycle is required for the L-
PBF and LP-DED 17–4 PH SS specimens to
diminish their variation in texture. Both the L-
PBF and LP-DED specimens had a twin-based
martensite with a K-S relationship to the parent
austenite.
5. The L-PBF 17–4 PH SS specimen outperformed
the LP-DED and wrought counterparts with the
same heat treatment in the room-temperature
tensile properties. This was attributed to the
finer microstructure of the L-PBF specimens.
6. The d-ferrite phase in the CA-H900 heat treated
LP-DED specimen reduces the ductility com-
pared with that of the L-PBF counterpart even
though the LP-DED specimen had lower tensile
strength. The high contrast between the Schmid
factor of the d-ferrite phase with the martensite
matrix in the {110}<111> slip system (i.e., high
incompatibility with matrix), as well as the
sharp edges of the d-ferrite phase, justified the
reduction in the ductility of the material.
7. The quasi-cleavage facets on the fracture sur-
faces of the CA-H900 treated specimens repre-
sented brittle behavior of the L-PBF and LP-
DED 17–4 PH SS in the CA-H900 condition,
while the fibrous fracture surface with dimples
depicted ductile behavior of the material in CA-
H1150 condition.
ACKNOWLEDGEMENTS
This paper is based upon the work partially fun-
ded by the National Aeronautics and Space
Administration (NASA) under Award
#80MSFC19C0010. Any subjective views or opin-
ions that might be expressed in the paper do not
necessarily represent the views of NASA or the
United States Government.
CONFLICT OF INTEREST
On behalf of all authors, the corresponding author
states that there is no conflict of interest.
Nezhadfar, Gradl, Shao, and Shamsaei
REFERENCES
1. N. Shamsaei, A. Yadollahi, L. Bian, and S.M. Thompson,
Addit. Manuf. 8, 12 (2015).
2. D.D. Gu, W. Meiners, K. Wissenbach, and R. Poprawe, Int.
Mater. Rev. 57, 133 (2013).
3. T.D. Ngo, A. Kashani, G. Imbalzano, K.T.Q. Nguyen, and D.
Hui, Compos. Part B 143, 172 (2018).
4. T.F. Babuska, B.A. Krick, D.F. Susan, and A.B. Kustas,
Manuf. Lett. 28, 30 (2021).
5. A. Jinoop, C. Paul, and K. Bindra, J. Mater. Des. Appl. 233,
2376 (2019).
6. R. Molaei, A. Fatemi, N. Sanaei, J. Pegues, N. Shamsaei, S.
Shao, P. Li, D.H. Warner, and N. Phan, Int. J. Fatigue 132, 1
(2020).
7. J. Simpson, J. Haley, C. Cramer, O. Shafer, A. Elliott, B.
Peter, L. Love, and R. Dehoff, ORNL/TM-2019-1190. Oak
Ridge Natl. Lab. 1190, 1 (2019).
8. Q. Guo, C. Zhao, M. Qu, L. Xiong, L.I. Escano, S.M.H. Hoj-
jatzadeh, N.D. Parab, K. Fezzaa, W. Everhart, T. Sun, and
L. Chen, Addit. Manuf. 28, 600 (2019).
9. A. Saboori, A. Aversa, G. Marchese, S. Biamino, M. Lom-
bardi, and P. Fino, Appl. Sci. 10, 3310 (2020).
10. M. Ma, Z. Wang, and X. Zeng, Mater. Sci. Eng. A 685, 265
(2017).
11. K.N. Amato, S.M. Gaytan, L.E. Murr, E. Martinez, P.W.
Shindo, J. Hernandez, S. Collins, and F. Medina, Acta Ma-
ter. 60, 2229 (2012).
12. F. Liu, X. Lin, H. Leng, J. Cao, Q. Liu, C. Huang, and W.
Huang, Opt. Laser Technol. 45, 330 (2013).
13. L.L. Parimi, G. Ravi, D. Clark, and M.M. Attallah, Mater.
Charact. 89, 102 (2014).
14. W. Xu, M. Brandt, S. Sun, J. Elambasseril, Q. Liu, K. La-
tham, K. Xia, and M. Qian, Acta Mater. 85, 74 (2015).
15. B.E. Carroll, T.A. Palmer, and A.M. Beese, Acta Mater. 87,
309 (2015).
16. J. Schneider, JOM 72, 1085 (2020).
17. C. Doñate-Buendia, R. Streubel, P. Kürnsteiner, M.B.
Wilms, F. Stern, J. Tenkamp, E. Bruder, S. Barcikowski, B.
Gault, K. Durst, J.H. Schleifenbaum, F. Walther, and B.
Gökce, Procedia CIRP 94, 41 (2020).
18. M.R. Stoudt, R.E. Ricker, E.A. Lass, and L.E. Levine, JOM
69, 506 (2017).
19. S. Cheruvathur, E.A. Lass, and C.E. Campbell, JOM 68, 930
(2015).
20. W. Liu, J. Ma, M.M. Atabaki, R. Pillai, B. Kumar, U.
Vasudevan, H. Sreshta, and R. Kovacevic, Lasers Manuf.
Mater. Process 2, 74 (2015).
21. S. Cao, Determination of the Fe-Cr-Ni and Fe-Cr-Mo phase
diagrams at intermediate temperatures using a novel dual-
anneal diffusion-multiple approach, Ph.D thesis, The Ohio
State University, 2013.
22. TCFE9 Thermo-Calc Software, 1 (2017).
23. P.D. Nezhadfar, E. Burford, K. Anderson-Wedge, B. Zhang,
S. Shao, S.R. Daniewicz, and N. Shamsaei, Int. J. Fatigue
123, 168 (2019).
24. P.D. Nezhadfar, P. Gradl, S. Shuai, and N. Shamsaei, In
Proc. 32nd Annu. Int. Solid Free. Fabr. Symp. – An Addit.
Manuf. Conf. (2021).
25. ASTM Standard E8/E8M-13a, Standard Test Methods for
Tension Testing of Metallic Materials (2006).
26. T. LeBrun, T. Nakamoto, K. Horikawa, and H. Kobayashi,
Mater. Des. 81, 44 (2015).
27. L. Facchini, N. Vicente, I. Lonardelli, E. Magalini, P. Ro-
botti, and A. Molinari, Adv. Eng. Mater. 12, 184 (2010).
28. M. Alnajjar, F. Christien, K. Wolski, and C. Bosch, Addit.
Manuf. 25, 187 (2019).
29. P.D. Nezhadfar, K. Anderson-Wedge, S.R. Daniewicz, N.
Phan, S. Shao, and N. Shamsaei, Addit. Manuf. 36, 101604
(2020).
30. R.K. Okagawa, R.D. Dixon, and D.L. Olson, Weld. Res.
Suppl. 62, 204s (1983).
31. P.D. Nezhadfar, R. Shrestha, N. Phan, and N. Shamsaei,
Int. J. Fatigue 124, 188 (2019).
Microstructure and Deformation Behavior of Additively Manufactured 17–4 Stainless Steel:
Laser Powder Bed Fusion vs. Laser Powder Directed Energy Deposition
32. C. Celada-Casero, J. Sietsma, and M.J. Santofimia, Mater. 37. ASTM A693-16, Standard specification for precipitation-
Des. 167, 107625 (2019). hardening stainless and heat-resisting steel plate, sheet, and
33. D.H. Ping, S.Q. Guo, M. Imura, X. Liu, T. Ohmura, M. strip (2016).
Ohnuma, X. Lu, T. Abe, and H. Onodera, Sci. Rep. 8, 14264 38. M.E. Stevenson, P.D. Umberger, and S.F. Uchneat, Fracture
(2018). appearance and mechanisms of deformation and fracture
34. C. Cayron, Acta Crystallogr. Sect. A 69, 498 (2012). (ASM International, 2021).
35. P.D. Nezhadfar, A. Zarei-Hanzaki, S.S. Sohn, and H.R.
Abedi, Mater. Sci. Eng. A 665, 10 (2016). Publisher’s Note Springer Nature remains neutral with re-
36. G.E. Dieter and D.J. Bacon, Mechanical metallurgy
gard to jurisdictional claims in published maps and institutional
(McGraw-Hill, New York, 1986).
affiliations.